Strike Force · Settings
Visual Shell
Prompt Registry
Visual Shell
+ New Prompt
DB-backed, versioned prompt templates. Every AI call records the
(prompt id, version) pair it ran against so
replay stays deterministic after edits.
| Prompt Id | Strategy | Active | Versions | Last Updated | Updated By | Notes | Actions |
|---|---|---|---|---|---|---|---|
ai.insight.validator.v1
|
StrikeForce | v4 | 4 | 2026-04-20 10:30 UTC | chevbryan@gmail.com | Tightened Disagree rationale to <= 2 sentences; schema now requires KeyRisks array. | Edit History |
fingerprint.narrative.v1
|
StrikeForce | v3 | 3 | 2026-04-20 07:30 UTC | chevbryan@gmail.com | Forbid contradicting structured sections; added drift-state context. | Edit History |
archetype.classifier.v1
|
StrikeForce | v2 | 2 | 2026-04-19 09:30 UTC | chevbryan@gmail.com | Added SympathyMover vs SectorMomentumSpillover tiebreak rule; schema unchanged. | Edit History |
catalyst.tagger.v1
|
StrikeForce | v6 | 6 | 2026-04-18 11:30 UTC | chevbryan@gmail.com | Dropped press-release subcategory split; Alpaca headline fallback stays. | Edit History |
global.common.system.v1
|
Shared | v1 | 1 | 2026-04-12 12:30 UTC | chevbryan@gmail.com | Shared framing: you are a deterministic trading-strategy assistant... | Edit History |
catalyst.tagger.experimental
Retired
|
StrikeForce | v2 | 2 | 2026-04-06 12:30 UTC | chevbryan@gmail.com | Shadow-only A/B variant. Retired until v0.10.1 learning-loop infra lands. | Edit History |